Y. Murata, N. Shibata, K. Yasumoto, and M. Ito (Japan)
Software Agents, Artificial Intelligence Applications, Modelling and Simulation
Efficiency of Genetic Algorithms (GAs) depends largely on the parameters such as crossover rate and mutation rate. In general, however, it is difficult to adjust those parameters manually. Although there are a few researches about adap tive GAs for adjusting multiple parameters, they require ex tremely large computation costs. In this paper, we propose a new algorithm based on multi agent techniques which combines existing meta-GA techniques and GAs with dis tributed environment schemes. Through some simulations, we have confirmed that the proposed algorithm can adapt multiple parameters in reasonable computation costs.
Important Links:
Go Back
